Difference between revisions of "User:Firespeaker/HFST bug"

From Apertium
Jump to navigation Jump to search
m (Freenode -> OFTC)
 
(11 intermediate revisions by 2 users not shown)
Line 1: Line 1:
{{TOCD}}
In 2011, a bug in how HFST handles words containing spaces was [http://sourceforge.net/p/hfst/bugs/59/ documented and resolved] (apparently in [http://hfst.svn.sourceforge.net/viewvc/hfst?view=revision&revision=1518 r1518]?), but it introduced a new bug. This page documents the new behaviour.
In 2011, a bug in how HFST handles words containing spaces was [http://sourceforge.net/p/hfst/bugs/59/ documented and resolved] (apparently in [http://hfst.svn.sourceforge.net/viewvc/hfst?view=revision&revision=1518 r1518]?), but it introduced a new bug. This page documents the new [incorrect!] behaviour. It appears to only affect transducers written in lexc.

[https://sourceforge.net/p/hfst/bugs/153/ A bug report] was filed in January of 2013 along with a patch for a test case. In March of 2013, [[User:Francis Tyers|spectie]] posted a patch that fixed the bug but introduced an issue with newlines and full stops. As of today, the bug has still not been fixed.


== text.lexc ==
== text.lexc ==
Line 35: Line 38:
* <code>$ echo "erke medvedev" | hfst-proc test2.hfst.ol</code>
* <code>$ echo "erke medvedev" | hfst-proc test2.hfst.ol</code>
: <code>^erke/erke$ ^medvedev/medvedev$</code>
: <code>^erke/erke$ ^medvedev/medvedev$</code>

== Another test case ==
This one is meant to be more familiar to English-speakers :)

<pre>
Multichar_Symbols

%

LEXICON Root

word:word #;
word% form:word% form #;
formation:formation #;
</pre>

* <code>$ echo "word formation" | hfst-proc test3.hfst.ol</code>
: <code>^word formation/<span style="color:red;">*</span>word formation$</code>
* <code>$ echo "formation word" | hfst-proc test3.hfst.ol</code>
: <code>^formation/formation$ ^word/word$</code>


== Notes ==
<ul>
<li>This doesn't seem to affect transducers written in other formats. E.g., the transducer that results from <code>apertium-eng-kaz.eng.dix</code> outputs the following:</li>
<ul>
<li><code>$ echo "right there" | apertium -d . eng-kaz-morph</code></li>
: <code>^right there/right there<adv>$^./.<sent>$</code>
<li><code>$ echo "right the" | apertium -d . eng-kaz-morph</code></li>
: <code>^right/right<adj>/right<adv>/right<n><sg>$ ^the/the<det><def><sp>$^./.<sent>$</code>
<li><code>$ echo "right therein" | apertium -d . eng-kaz-morph</code></li>
: <code>^right/right<adj>/right<adv>/right<n><sg>$ ^therein/*therein$^./.<sent>$</code>
</ul>
</ul>


== Other materials ==
== Other materials ==
* [http://wiki.apertium.org/wiki/Talk:Ideas_for_Google_Summer_of_Code/Closer_integration_with_HFST spectie explains the bug to firespeaker]
* [http://wiki.apertium.org/wiki/Talk:Ideas_for_Google_Summer_of_Code/Closer_integration_with_HFST spectie explains the bug to firespeaker]
* irc.oftc.net#hfst

Latest revision as of 06:33, 27 May 2021

In 2011, a bug in how HFST handles words containing spaces was documented and resolved (apparently in r1518?), but it introduced a new bug. This page documents the new [incorrect!] behaviour. It appears to only affect transducers written in lexc.

A bug report was filed in January of 2013 along with a patch for a test case. In March of 2013, spectie posted a patch that fixed the bug but introduced an issue with newlines and full stops. As of today, the bug has still not been fixed.

text.lexc[edit]

Make sure to include the space in '% ' under Multichar_Symbols.

Multichar_Symbols

% 

LEXICON Root

erke:erke # ;
erke% me:erke% me # ;
medvedev:medvedev # ;

Compiling[edit]

  1. $ hfst-lexc test.lexc -o test.hfst
  2. $ hfst-invert test.hfst | hfst-fst2fst -w -o test.hfst.ol

Testing[edit]

Some correctly analysed forms[edit]

  • $ echo "erke" | hfst-proc test.hfst.ol
^erke/erke$
  • $ echo "erke me" | hfst-proc test.hfst.ol
^erke me/erke me$
  • $ echo "medvedev" | hfst-proc test.hfst.ol
^medvedev/medvedev$

The incorrectly analysed form[edit]

  • $ echo "erke medvedev" | hfst-proc test.hfst.ol
^erke medvedev/*erke medvedev$

Expected output[edit]

This form is analysed correctly by a transducer identical to the one above except with the "erke me" form removed:

  • $ echo "erke medvedev" | hfst-proc test2.hfst.ol
^erke/erke$ ^medvedev/medvedev$

Another test case[edit]

This one is meant to be more familiar to English-speakers :)

Multichar_Symbols

% 

LEXICON Root

word:word #;
word% form:word% form #;
formation:formation #;
  • $ echo "word formation" | hfst-proc test3.hfst.ol
^word formation/*word formation$
  • $ echo "formation word" | hfst-proc test3.hfst.ol
^formation/formation$ ^word/word$


Notes[edit]

  • This doesn't seem to affect transducers written in other formats. E.g., the transducer that results from apertium-eng-kaz.eng.dix outputs the following:
    • $ echo "right there" | apertium -d . eng-kaz-morph
    • ^right there/right there<adv>$^./.<sent>$
    • $ echo "right the" | apertium -d . eng-kaz-morph
    • ^right/right<adj>/right<adv>/right<n><sg>$ ^the/the<det><def><sp>$^./.<sent>$
    • $ echo "right therein" | apertium -d . eng-kaz-morph
    • ^right/right<adj>/right<adv>/right<n><sg>$ ^therein/*therein$^./.<sent>$

Other materials[edit]